Barton, K. E., & Huynh, H. (2000, April). Patterns of errors made on a reading test with oral reading administration . Annual meeting of the National Council on Measurement in Education (NCME), New Orleans, LA, United States.

All students were read the test by a reader (read-aloud), by an audio recording operated by either a test administrator or by the student, or by a video recording of a signed administration for students with hearing impairments. All test directions, items, and answer choices were read (or signed) to the students.

Score data from high school students with physical (n=91), emotional (n=125), and learning disabilities (n=1902), students with intellectual disabilities (n=438), and 368 students with no disabilities from South Carolina (U.S.) were analyzed. The physical disability category included students with hearing, orthopedic, vision, and language impairments. All students had completed the South Carolina Basic Skills Assessment Program (BSAP) High School Examination of Reading, Oral Accommodation form. Of the total sample, 87% of the students were students with disabilities. 36% of the students were female, and 64% were male. The dataset originally included 9 Asians, 1674 African Americans, 12 Hispanics, 15 American Indians, and 1289 White students; however, the Asian, Hispanic, and American Indian students were excluded from the analysis due to small population sizes, as well as students who each had two or three listed disabilities (40 students).

A high school exit exam multiple choice test, more specifically the 1996-1998 South Carolina Basic Skills Assessment Program (BSAP) High School Examination Test of Reading, oral accommodation form, was used as the dependent variable. Items were analyzed to determine what types of errors were associated with the read aloud accommodation form for students with different types of disabilities.

Associations were identified by disability groups for errors made, as well as level of achievement. Overall, the items that were characteristic of literal comprehension found in specific detail or reference questions, but not those examining to inferential comprehension—such as in main idea questions—appeared to have the strongest association with disability status.